134 REPORT— 1882. 



for the distance between a and y and between /3 and e is actaally larger 

 in the chloride than in the bromide, contrary to what holds for the re- 

 maining bands. In the case of barium iodide, Mitscherlich's rule cannot 

 be said to hold. 



A more satisfactory agreement could be obtained if the distances 

 between the corresponding bands were made to depend, not on the 

 atomic weights of the compound, but on the atomic weights of the 

 metalloid in the compound ; but no object is gained at present by giving 

 such rules before sufficient material has been accumulated to prove or 

 disprove them. The second part of Mitscherlich's rule, which regulates 

 the displacement of the bands in the different compounds, does not stand 

 very well when tested by Lecoq's measurements ; the band /3 of barium 

 iodide, for instance, ought, according to it, to be at 5356 instead of 5376. 



We ought to inquire whether a connection can be traced between the 

 lines of a metal and that of one of its compounds. It would be possible, 

 for instance, that the oxide should show its bands chiefly at such places 

 at which we find lines in the metallic spectrum, and such a rule might 

 be suggested by the examination of the calcium spectrum, which shows a 

 characteristic group of lines exactly at the place which is filled by the 

 green band of the oxide. No general rule can, however, be given, and in 

 some cases even the metallic spectrum seems particularly free of lines in or 

 about the place in which we find the oxide bands. There is at present no 

 hope whatever of directly connecting the spectrum of a metal and that of 

 its compounds, though, as was seen, we may hope to gain an insight into 

 the relations of the spectra of such similar compounds as the chlorides, 

 bromides, and iodides, which may be supposed to have a similar con- 

 stitution. 



Some very interesting changes have been noticed in the position of 

 absorption bands when certain colouring matters are dissolved in different 

 liquids. We mention the names only of Hagenbach, Kraus, and Kundfc 

 as having studied the question in particular cases. Two papers by 

 Kundt ' and Claes ^ contain all that is known at present on the subject, 

 and it will be sufficient therefore to consider them only. Kundt 

 examined the position of the absorption bands of chlorophyll, anilingreen, 

 cyanin, fuchsin, chinizarin, and of the colouring matter of the yolk of 

 egg when dissolved in a number of liquids. His results are given in 

 Table VIII., which is taken out of his paper. The solvents are arranged 

 in each column in such order that with the one first on the list the 

 absorption band is seen most towards the blue, while the one which 

 appears last disjDlaces the band most towards the red. 



In the dissertation published by Dr. Claes we must carefully dis- 

 tinguish the experimental from the theoretical part. We have at present 

 no reason to doubt the accuracy of his experimental work, and as he 

 used solvents and liquids partly diffei-ent from those employed by Pro- 

 fessor Kundt, we give his results in Table IX. arranged in the same way 

 as Table YIII. We give the results for chlorophyll in two columns ; one 

 refers to the absorption baud in the red, the second to the absorption 

 bands in the green. Magdala red shows different absorption bands in 

 two different sets of solutions, and two columns are therefore also 

 necessary. The results obtained with different didymium compounds are 

 not tabulated, as they are partly due to a different order of phenomena. 



■ Wied. Ann. iv. p. 34 (1878). » Wied. Ann. iii. p. 389 (1878). 



